High Application Drop-Offs
Customers abandoned applications partway through when they hit a validation error they didn't understand, with no real-time assistance to guide them back on track.
Step-by-step spoken audio instructions embedded directly into the app journey
Bilingual voice support enabling users to navigate forms and processes seamlessly
Instant contextual prompts that detect user errors and guide accurate completion
About the client
Bajaj Finserv (Bajaj Finance) offers Personal Loan, Business Loan, and Home Loan applications through a "3-in-1" consumer app used by a large and diverse customer base, including Tier 2 and Tier 3 city users.
Loan application forms carry strict validation, PAN, date of birth, pincode, and consent checkboxes among them, which can be confusing for customers who hit an error without understanding what went wrong or how to fix it.
The Challenge
Customers abandoned applications partway through when they hit a validation error they didn't understand, with no real-time assistance to guide them back on track.
A field would turn invalid without customers noticing which one or why, leading them to give up rather than hunt for the problem.
Tier 2 and Tier 3 city customers, more comfortable with spoken Hindi than written English instructions, struggled to complete applications on their own.
Without field-level instrumentation, BFL's team couldn't see which specific fields were driving the most abandonment.
How Gyde solved it
When a customer tries to submit a form that still has errors, Gyde Voice Assist detects the failed attempt and offers help, letting the customer choose English or Hindi.
The assistant highlights the first invalid field and plays a short spoken instruction explaining exactly what to do, advancing to the next problem field as each one is corrected.
The same experience runs across BFL's Personal Loan, Business Loan, and Home Loan applications, supporting BFL's custom form components across web and Android.
A dependency-free web widget covers BFL's progressive web app, and a native Android SDK covers the mobile app, both drawing from the same authored voice "flows" and both feeding session and interaction data back for analytics.
Why It Works
A dependency-free web widget and documented Android SDK connect to existing loan application systems without major engineering rework.
The Android SDK includes root/compromised-device detection, encrypted token storage, and certificate-pinned network calls to protect customer data.
Gyde's team authors and updates voice guidance per product and language directly in the dashboard, with no engineering dependency to change what customers hear.
Explore more success stories
Give your customers real-time, spoken guidance in the moment they get stuck, in the language they're comfortable with.